This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
[RFC] 12266 Fallout
- From: Keith Seitz <keiths at redhat dot com>
- To: gdb-patches at sourceware dot org
- Date: Thu, 25 Aug 2011 16:02:59 -0700
- Subject: [RFC] 12266 Fallout
Hi,
I am writing for some maintainer guidance on an issue that has arisen.
The original reporter for c++/12266 reports that the bug is not fixed.
And indeed, it isn't!
The reporter's test case (C++):
typedef std::string foo;
void calltest (foo) {}
Setting a break at "calltest(foo)" does not work because inspect_type
(part of 12266 patchset) uses check_typedef, which resolves foo ->
std::string -> std::basic_string<char, std::char_traits<char>,
std::alloactor<char> >. But because we are no longer using DMGL_VERBOSE
in dwarf2_physname (and other places) to do demangling, the symbol table
actually stores "calltest(std::string)".
Setting a break at "calltest(std::string)" works.
Setting a break at "calltest(std::basic_string<char, ...>)" also does
not work for the same reason "foo" didn't work.
There are two ways I can see to fix this.
1) Add DMGL_VERBOSE where it is needed so that NO typedefs ever appear
in a (non-TYPE_CODE_TYPEDEF) symbol's name, i.e, store
"calltest(std::basic_string<...>)" in the symbol table instead of
"calltest(std::string)".
2) Add a new version of check_typedef which "stops" at std::string (and
std::ostream, std::istream, std::iostream), and add logic (somewhere) to
deal with going from "std::basic_string<...>" back to "std::string" (and
likewise for the others).
For whatever it may be worth, my preference is for #1. I've always felt
storing the most fundamental representation of a symbol was the best
option: it treats all symbols the same.
Along these lines, I have done a quick audit of the code (look for
DMGL_PARAMS), and the only places where DMGL_VERBOSE needs to be added
is dwarf2_physname, lookup_symbol_in_language, and
symbol_find_demangled_name. All other users of DMGL_PARAMS are using it
for output to the display, and leaving out DMGL_VERBOSE would be preferred.
This will have the immediate side effect of printing symbol names with
"std::basic_string<...>" instead of the more preferred "std::string".
Actually in this case, the "right" output would be "calltest(foo)", but
that is a different, but related, problem, for which I developed a patch
a long time ago, part of the first 12266 patch submissions
(dwarf2_print_name).
So, I humbly ask maintainers:
- Shall I continue #1 and start submitting patches?
- Shall I start teaching gdb about std::string et al and how to deal
with them in the environment we have today?
- Have I missed something that would be preferable to anything I've
mentioned?
Keith